Nematode germ granule assembly is linked to mRNA repression

RNA-protein (RNP) granules are non-membrane bound organelles with enigmatic roles in RNA metabolism. Metazoa contain RNP germ granules specialized for germline development. Caenorhabditis elegans P-granules are liquid droplet germ granules that require PGL proteins for assembly. Here we investigate PGL proteins to understand the relationship between P-granule assembly and germline function. We determine the crystal structure of a PGL N-terminal domain (NTD) and find that it dimerizes. From the structure, we identify mutations that disrupt PGL dimerization in vitro and prevent PGL granule formation in mammalian cells in culture. These same mutations in nematodes prevent assembly of PGL into P-granules and cause sterility. Using a protein-mRNA tethering assay, we show that mRNAs recruited to PGL-1 are repressed, while mRNAs recruited to PGL-1 mutants defective for granule assembly are expressed. Therefore, the effects of PGL on mRNA repression and fertility are tightly linked to its formation of higher-ordered assemblies.


40
Subcellular localization can be critical for RNA control. The locations of RNAs and RNA 41 regulatory proteins within a cell can dictate whether RNAs are translated or repressed (Singh et sequestration and translation and have been proposed to function in mRNA regulation (Buchan, 5 role in PGL function. Our initial efforts to express trypsin-mapped recombinant protein fragments at the nuclear periphery ( Figure 3D), similar to those seen with antibody staining to untagged 147 PGL-1 and PGL-3 (Kawasaki et al., 2004;Kawasaki et al., 1998). The SNAP-tagged protein 148 therefore provides a simple way to evaluate PGL assembly into granules.

150
To probe the role of PGL NTD dimerization in the nematode germline, we introduced the 151 assembly mutants into PGL-1::SNAP ( Figure S1B) and assayed effects on fertility and granule 152 formation. For both K126E K129E and R123E mutants, many were sterile at 20°C and nearly all 153 were sterile at 25°C ( Figure 3B). Indeed, the percentage of sterile animals was higher than 154 seen in a pgl-1 null mutant (Figure 3B), suggesting that abolishing NTD dimerization had a 7 dominant-negative effect. Both interface mutants had smaller than normal germlines and many 156 lacked oocytes (Figure S3B-D), similar to pgl-1 and pgl-1 pgl-3 null mutant germlines S3). Both K126E K129E and R123E mutant proteins were expressed, but their distribution was 160 largely diffuse (Figures 3E-F  capable of incorporating into P-granules, but do so much more weakly than their wild-type 165 counterparts ( Figure 3D). We conclude that PGL NTD dimerization is critical for fertility and 166 efficient PGL granule formation.

Granular PGL represses mRNAs in vivo
assay, widely used to investigate RNA regulatory proteins (Baron-Benhamou et al., 2004;Coller 194 and Wickens, 2002). Our assay examined the fate of mRNAs to which PGL-1 was tethered via 195 λN22, a short peptide that binds with high affinity and sequence specificity to the boxB RNA 196 hairpin ( Baron-Benhamou et al., 2004). Versions of this method were used previously in worms 197 and other organisms (Baron-Benhamou et al., 2004;Wedeles et al., 2013). For the reporter, we 198 inserted three boxB sites into the 3'UTR of a ubiquitously-expressed, germline GFP-histone 199 reporter ( Figure 4A, Methods) (Zeiser et al., 2011). To tether PGL to the GFP reporter mRNA,

212
The wild-type PGL-1::SNAP::λN22 formed perinuclear granules ( Figure 4E-F), though the 213 SNAP signal was lower, perhaps because animals were heterozygous. Regardless, the key 214 conclusion is that PGL-1 tethering dramatically decreased GFP expression from the mRNA 215 reporter.

217
One possible explanation for loss of GFP expression might have been germline silencing, a 218 phenomenon common for genes expressing foreign proteins and thought to prevent deleterious 219 mRNAs from entering the cytoplasm (Hoogstrate et al., 2014). To ask if the reporter had been 220 silenced, we used single molecule fluorescence in situ hybridization (smFISH) to detect gfp nuclear and cytoplasmic puncta ( Figure S4D). We interpret the nuclear puncta as active 223 transcription sites and cytoplasmic puncta as mRNAs, based on a previous study (Lee et al., 224 2016). In these control germ cells, GFP protein fluorescence was robust ( Figure S4B), and 225 PGL-1::SNAP localized to perinuclear granules ( Figure S4C). Germ cells harboring PGL-226 1::SNAP::λN22 ( Figure S4E-H,M) also possessed nuclear and cytoplasmic puncta ( Figure   227 S4H,M), but cytoplasmic puncta were fewer and frequently colocalized with P-granules ( Figure   228 S4G-H,M). These germ cells possessed gfp reporter transcripts and thus were not subject to germline silencing. However, they had no GFP protein expression, indicating that reporter expression was repressed by PGL-1 tethering.

10
Our structural insight into PGL provided an opportunity to test how PGL granule formation 233 affected its ability to repress mRNAs. We introduced K126E K129E into PGL-1::SNAP::λN22 to 234 prevent robust granule assembly (Figure S1B). Mutant homozygotes had modest fertility (21% 235 fertile, n=96) that was comparable to other NTD mutant worms ( Figure 3B). Dimerization-236 defective PGL-1 failed to repress the reporter RNA to which it was tethered: most germ cells    2016). In this work, we report the discovery of PGL NTD dimerization and demonstrate that PGL 254 NTD dimerization is critical for granule formation, fertility and mRNA repression in vivo. Based 255 on these results, we propose that mRNAs in P-granules are repressed and that this repression 256 requires PGL assembly into granules ( Figure 4H). PGL contains two dimerization domains (this 11 future direction is to investigate how each dimerization domain contributes to higher order and 259 likely oligomeric assembly.

261
Granule assembly proteins form a structural network that relies on multivalency and low affinity 262 interactions for plasticity (Bergeron-Sandoval et al., 2016). We have discovered that PGL uses at least one dimerization domain to form granules, but our results do not address the low affinity 264 interactions that must be present to drive liquid droplet behavior. Critical granule assembly  (Decker et al., 2007;Jeske et al., 2015;Ling et al., 2008;Nott et al., 2015; 271 Wang et al., 2014). We suggest that PGL also makes low affinity contacts that work with its

280
Our study provides direct evidence that recruiting mRNAs to a liquid droplet RNP granule                              20 Figure 3. NTD dimerization is critical for fertility and P-granule formation in nematodes region of meiotic pachytene (see Figure S3A). (C-F) Representative images of SNAP staining 477 to visualize PGL-1 expression and granule formation. All images are partial z-stacks to 478 maximize visualization of P-granules. Images were taken from germlines producing embryos; 479 similar images were obtained from germlines too defective to make embryos (Figure S3F-G).

480
(C) Control, wild type animal lacking SNAP tag shows virtually no background staining (n=20).

116
injection marker phenotype were additionally screened by a combination of PCR without or with restriction 117 enzyme digest to identify those with the repair of interest. In JK5687, a SNAP tag (Keppler et al., 2003) 118 was inserted between PGL-1 amino acids G713 and G714 in N2 worms. A 3xMYC tag was added to the 119 N-terminus of GLH-1 between G17 and F18. A 3xV5 tag was added in the C-terminal region of PGL-3 120 between residues G627 and S628. F2s were PCR screened to identify homozygous SNAP alleles and 121 the PCR product sequenced to confirm proper repair. Three worm strains were too infertile to freeze. All 122 worms were outcrossed at least twice with N2, with the exception of (glh-1(q858)[GLH-

126
Worms were singled into the peripheral wells of a 24-well plate that contained NGM agar and OP50 127 bacteria. Worms were allowed to propagate for 5 days at 20°C or 25°C, and then scored for progeny and 128 gravid progeny. We report the progeny numbers here.

130
To analyze GFP reporter expression, L4 larvae were propagated for approximately 24 hours at 20°C, 131 placed in M9 with 0.1 mM levamisole on a glass slide with a cover slip, imaged at 10x magnification on a 132 compound microscope and counted for the presence or absence of GFP fluorescence in its germline.

157
Samples were imaged using a Leica SP8 scanning laser confocal microscope, taking 0.3 µm (smFISH 158 experiments) or 1 µm (protein staining) slices in sequence. Maximum intensity partial stack projections 159 were generated and brightness adjusted using ImageJ (Schindelin et al., 2015). All images were treated 160 equally in ImageJ and Photoshop, with the exception of the transmitted light images. Imaging directed genome editing in Caenorhabditis elegans using CRISPR-Cas9 ribonucleoprotein